Open AccessSuperpixel Image Segmentation Based on Improved K-means
Author(s) -
Ran Siyuan,
Xinying Li
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1533/3/032067
Subject(s) - artificial intelligence , segmentation , computer science , image segmentation , scale space segmentation , segmentation based object categorization , pattern recognition (psychology) , similarity (geometry) , computer vision , minimum spanning tree based segmentation , image (mathematics)
Superpixel algorithm is the pretreatment of the image processing steps, in the processing of computer vision, precision of the superpixel algorithm is very important. In order to improve the accuracy of superpixel segmentation, an improved k-means algorithm for superpixel segmentation is proposed. This algorithm uses the improved k-means algorithm to cluster the images. By changing the calculation method of distance, it improves the measurement of similarity, improves the accuracy of segmentation, and improves the quality of segmentation results. Experimental data show that this algorithm performs better than the traditional algorithm and improves the segmentation quality and visual effect.